Personally, I do not give two identical fucks in the world for the OpenAI fiasco. It’s a bunch of AI tech bros who feel that laws do not apply to their “scientific” research. The ensuing FUD is being contributed by those who can’t be bothered to do deeper research further convoluting the facts. You breach companies, you are liable and held to the CFAA. Aaron Swartz, if you don’t know him then look him up, was arrested by MIT police and was facing serious federal charges under the Computer Fraud and Abuse Act for downloading JSTOR publications from MIT under his guest account. He felt that knowledge and scientific research should not be gated, and freely accessible by anyone. I agree with him and I support that ideal of freely accessible scientific research and publications. Human knowledge benefiting humans should not be paywalled or gated. OpenAI, however, knowingly, willingly, and admittedly removed safety guardrails, leveraged a misconfigured sandbox with known security issues, and didn’t say anything about it for days. OpenAI, fully understanding the risks that were discovered in third-party risk research they participated in and published by METR a few months prior to the breach, still decided to continue that model’s testing. In a weird way, the results of these breaches and the order of operations are almost identical to METR’s Risk Report, INC-026 through INC-030 here https://lnkd.in/gC3xSeNS OpenAI and the like should be charged with CFAA. The situation here is arguably worse and shouldn’t be dismissed just because “AI” is prepended to the incident. Stop treating this like an AI problem. It is still a computer. Someone still needed to operate the system. Someone is still liable. Aaron didn’t deserve it, but OpenAI does.